Background

H. pylori is a Gram-negative bacterium that colonizes the human gastric epithelium. H. pylori infection is characterized by mucosal inflammation (gastritis) and may result in peptic ulcer disease or gastric adenocarcinoma [1]. H. pylori is transmitted between humans by a variety of routes including gastro-oral and fecal-oral mechanisms that include contaminated water and food. It has also been postulated that oral cavity may play a role in H. pylori transmission and possibly act as a reservoir [2]. For example, H. pylori has been detected in the oral cavity using the polymerase chain reaction (PCR). H. pylori has also been successfully cultured from saliva from individuals with positive results of both saliva H. pylori antigen test and H. pylori flagella test [3].

The oral cavity is one of the most complex and largest microbial habitats that harbors hundreds of different bacteria which play important roles in maintaining oral homeostasis and influencing the development of both oral and systematic diseases [4]. Many factors in the oral environment including intraoral pH and salivary iron concentration as well as expression of serum malondialdehyde and lipid profile have been reported to have significant relationships with oral microbial communities [5, 6]. Several recent studies detected the efficacy of antibiotics or drug composed of herbal on oral microbiol communities [7, 8]. However, there was few reports about H. pylori and its relationship with the microbial community structure in human saliva.

Currently, most oral bacteria species cannot be easily cultivated in vitro using traditional cultivation methods requiring the use of molecular biological techniques, such as checkerboard hybridization, microarray chips, and the quantitative real-time PCR to identify and classify the currently uncultivable bacteria [9]. However, many low-abundance bacteria species still cannot be detected using these approaches which impede the comprehensive and in-depth understanding of oral bacteria diversity. In this study, we used amplicon sequencing of 16S rDNA V3-V4 hypervariable regions to define the bacterial composition, abundance, and structure of salivary microbiome in people with and without active H. pylori infections. In addition, we also characterized the salivary biodiversity of a subgroup of subjects before and after the H. pylori eradication. As H. pylori passes through saliva, it is important to detect its expression in oral cavity as well as the efficacy of its infection and eradication on the oral microbiol communities.

Methods

Subjects and sample collections

The samples were collected in Renji Hospital of Shanghai Jiao Tong University, China, from August to November in 2018. A total of 58 subjects were recruited, including 34 subjects with H. pylori infection and 24 uninfected subjects. We first conducted a cross-sectional study of the salivary microbiota of 34 H. pylori infected and 24 uninfected subjects. A prospective study was then performed in a subgroup of 22 subjects with H. pylori infection who underwent salivary analysis both before and after successful H. pylori eradication.

All subjects received both endoscopy and 13C urea breath test (13C-UBT) before enrollment. The H. pylori infection status was confirmed by positive rapid urease test (RUT), histology and 13C-UBT. Absence of infection was defined as negative results for all tests (i.e., RUT, histology and 13C-UBT). H. pylori infected subjects received eradiation therapy consisting of esomeprazole 20 mg b.i.d., bismuth potassium citrate 600 mg b.i.d, amoxicillin 1000 mg t.i.d., and metronidazole 400 mg t.i.d. for 14 days. H. pylori eradication was confirmed by 13C-UBT at least 6 weeks after the end of treatment. Saliva samples were collected from 22 subjects both before and 2 months following successful H. pylori eradication. Subjects were characterized into four groups. H. pylori uninfected group (uninfected) (n = 24), H. pylori infected group (infected) (n = 34), and Pre-eradicated H. pylori infected group (pre-eradicated) (n = 22) and a successful eradicated group (eradicated) (n = 22). Inclusion criteria of subjects were: age of 20–65 years old male or female, with good oral hygiene (including brushing teeth twice a day for three minutes each) and with no bad eating habits [10]. Exclusion criteria included: 1) the presence of dental carious or any untreated cavitated carious lesions and oral abscesses, 2) periodontal disease or periodontal pockets ≥4 mm, 3) the use of antibiotics or proton pump inhibitors (PPIs) within 2 months before the study, 4) previous diagnosis of a serious systemic diseases (such as diabetes, hypertension or cardiopathy) or any diseases affecting oral health (such as Sjogren’s syndrome or any disease characterized by xerostomia), 5) pregnancy of breastfeeding, 6) specific eating habits or habits of eating specific food such as garlic or pepper and 7) smoking or alcohol drinking. The periodontal examinations were conducted by the same professional dentist in Renji Hospital of Shanghai Jiao Tong University, China, using the same diagnostic criteria. The detailed clinical parameters of the 58 subjects are shown in Additional file 1:Table S1.

Salivary sampling

Sampling was performed in the morning 2 h after eating. Saliva samples were collected from each subject according to the Manual of Procedures for Human Microbiome Project (http://hmpdacc.org/resources/tools_protocols.php), with minor modifications. Approximately 3–4 ml of non-stimulated saliva was collected in two sterile, labeled 2 mL Eppendorf tubes, which were immediately placed on ice. Within 3 h of collection, samples were transported on ice and stored at − 80 °C until use.

DNA extraction and sequencing

DNA was extracted from the saliva samples using the E.Z.N.A.® Soil DNA Kit (OMEGA, USA), following the manufacturer’s instructions, and stored at − 20 °C prior to further analysis. PCR amplification of the bacterial 16S rDNA hypervariable V3-V4 region was performed using the forward primer 338F (5′-ACTCCTACGGGAGGCAGCA-3′) and the reverse primer 806R (5′-GGACTACHVGGGTWTCTAAT-3′). Sample-specific 7-bp barcodes were incorporated into the primers for multiplex sequencing. Details of the barcodes are shown in Additional file 2: Table S2. PCR amplification were performed on an ABI 2720 instrument (ABI, USA) with an initial denaturation at 98 °C for 2 min, followed by 25 cycles of denaturation (15 s at 98 °C), annealing (30s at 55 °C), extension (30s at 72 °C), and ended with a final extension (5 min at 72 °C). A negative control was set during the process of PCR amplification to eliminate reagent contamination. PCR amplicons were purified with Agencourt AMPure Beads (Beckman Coulter, Indianapolis, IN) and quantified using the Quant-iT PicoGreen dsDNA Assay Kit (Invitrogen, Carlsbad, CA, USA). Equimolar concentrations of purified amplicons were pooled in equal amounts. Subsequently, the paired-end 2 × 300 bp sequencing plus 20% PhiX control DNA was performed on the Illumina MiSeq platform with MiSeq Reagent Kit v3 (Illumina, USA), following the vendor’s standard protocols.

Sequence analysis

The Quantitative Insights Into Microbial Ecology (QIIME) pipeline was employed to process the sequencing data. Raw sequences were filtered to obtain high-quality sequences according to QIIME [11]. The low-quality sequences were filtered through following criteria: sequences that had a length of < 200 bp, sequences that had average Phred scores of < 25, sequences that contained ambiguous bases or homopolymeric stretches, and sequences that contained mononucleotide repeats of > 8 bp. The high-quality sequences were clustered into operational taxonomic units (OTUs) at 97% sequence identity by UCLUST [12]. The representative sequences selected from each OTU were classified taxonomically by BLAST searching against the Human Oral Microbiome Database (HOMD). OTU taxonomic classification was conducted by BLAST searching the representative sequences set against the HOMD using the best hit, with a BLAST e-value <1e-5. HOMD provides a detailed record of the type, metabolism, and pathogenicity of oral bacteria [13]. Then, an OTU table was further generated to record the OTU abundance of each sample and the taxonomic classification of these OTUs. Finally, to minimize the difference of sequencing depth across samples, the OTU table was modified by removing OTUs containing less than 0.001% of total sequences across all samples for further analysis [14].

Bioinformatics and statistical analysis

Sequence data analyses were mainly performed using QIIME (version 1.8.0) and R packages (version 3.2.4). The alpha diversity analysis including Chao 1 richness estimator, Abundance-based Coverage Estimator (ACE) metric, Shannon diversity index, and Simpson index, were calculated at 97% identity using the QIIME [15]. Ranked abundance curves were generated to compare both the richness and evenness of OTUs among samples. The beta diversity analysis including Nonmetric Multidimensional Scaling (NMDS), and unweighted UniFrac distances based principal coordinate analysis (PCoA), were performed using the R package to evaluate the similarity among various bacterial communities [16]. The significance of differentiation of microbiota structure among groups was assessed by Adonis test [17]. The taxonomy compositions and abundances were visualized by MEGAN (version 6.6.7) software [18]. Linear discriminant analysis effect size (LEfSe) was used to compare the bacterial community structures between the samples from the patients with and without H. pylori infection, as well as before and after the eradication regimen, using the online Galaxy workflow framework (http://huttenhower.sph.harvard.edu/galaxy/) [19].

Results

Global sequencing data

The saliva samples were obtained from 34 H. pylori infected, 24 uninfected and 22 in 34 subjects 2 months after successful eradication therapy. A total of 2,812,659 high-quality sequences (representing 79% of the total sequences) were acquired from the 80 saliva samples, with an average of 35,158 sequences per sample (ranging from 19,210 to 44,310; Additional file 3: Table S3). The average sequence length was 445 bp, with the maximum length being 548 bp and the shortest length being 136 bp (Additional file 4: Figure S1). Clustering of all high-quality sequences at 97% identity by UCLUST resulted in 70,489 OTUs with an average of 4222 OTUs (ranging from 2170 to 5669), which were BLAST-searched against the HOMD for taxonomic assignments. After removing the low-credibility OTUs (together contributing only 6.7% of all sequences) and subsampling each sample to an equal sequencing depth of 19,210 reads per sample, a modified OTU table was obtained consisting of 95,812 OTUs with an average of 1198 OTUs per sample (ranging from 697 to 1584; Additional file 5: Table S4).

Bacterial abundance and distribution

The bacterial distribution was characterized in terms of the relative taxonomic abundances. A total of 11 phyla, 21 classes, 36 orders, 68 families, 138 genera and 440 species were detected in the saliva samples. The taxonomic distributions of the predominant bacteria (relative abundance > 1% of the total sequences) in subjects with and without H. pylori at different levels were shown in Fig. 1. The 6 most abundant phyla were Proteobacteria (40.1% of the total sequences), Firmicutes (31.6%), Bacteroidetes (13.0%), Actinobacteria (7.4%), Fusobacteria (6.1%), and TM7 (1.0%), together accounting for 99.2% of the total sequences. At genus level, saliva microbiota was dominated by Neisseria, Streptococcus, Haemophilus, Veillonella, and Prevotella, with average relative abundances of 20.2, 16.5, 10.5, 8.0, and 8.0%, respectively. Fusobacteria was observed higher abundance in eradicated group on phyla level. Leptotrichia, Campylobacter and Pseucomonas expressed higher in eradicated group on genus level, while Alloprevotella and Aggregatibacter were observed higher abundance in uninfected group. The compositions in taxa of the microbial communities according to the tested sample groupings are provided in Additional file 6: Figure S2.

Fig. 1
figure 1

Distribution of the predominant bacteria at different taxonomic levels (phylum, class, order, family, and genus). The predominant taxa (> 1% relative abundance) in each level are shown. Red represent P group, green represent N group, and blue represent E group. N = uninfected group, P = infected group, E = eradicated group

Full size image

Bacterial diversity analysis

The salivary microbiota richness, measured by numbers of observed OTUs, was similar in uninfected subjects and infected subjects (Additional file 7: Figure S3A). The alpha diversity indices of Chao1, ACE, Shannon, Inverse Simpson, Good’s coverage, and Simpson’s evenness are shown in Table 1. The Shannon diversity index was higher in uninfected subjects than in infected subjects, but there was no significant difference between groups by t-test (1417.58 vs. 1393.60, p > 0.05). Besides, the ACE richness index (1491.22 vs. 1465.97, p > 0.05), Chao 1 richness estimator (1417.58 vs. 1393.60, p > 0.05), and the Inverse Simpson diversity index (1.02 vs. 1.02, p > 0.05) was also higher in uninfected subjects, with no significant difference, indicating the similar bacterial diversity of H. pylori uninfected saliva compared to the infected subjects. Good’s coverage estimator for each group was over 98%, indicating that the current sequencing depth was sufficient to saturate the bacterial diversity of saliva. In addition, Simpson’s evenness index indicated that the bacterial-community distribution in two groups was uneven, which was also observed in the rank-abundance curve (Additional file 8: Figure S4).

Table 1 Alpha diversity indices for saliva bacteria in each group at 97% identity
Full size table

Bacterial community structures

To gain insights into the similarities in the bacterial community structures among uninfected and infected subjects, PCoA of beta diversity analysis was performed based on the unweighted UniFrac distances, which demonstrated different community structures among two groups (PERMANOVAR, pseudo-F: 1.49, p = 0.033). As shown in Fig. 2a, the overall microbial composition of infected subjects deviated from that of uninfected subjects. Furthermore, the results of NMDS based on the genus level classification exhibited clear segregations in community structures among groups (Additional file 9: Figure S5A).

Fig. 2
figure 2

Principal coordinate analysis (PCoA) of unweighted UniFrac analysis. a. PCoA analysis demonstrated that subjects of P group were significantly different from N group (PERMANOVAR, pseudo-F: 1.49, p = 0.033). N = uninfected group, P = infected group. b. PCoA analysis showed that the overall microbial composition showed significant difference between PE and E group (PERMANOVAR, pseudo-F: 3.34, p = 0.001). E = eradicated group, N = uninfected group, PE = pre-eradicated group

Full size image

Differential microbiota compositions

There were significant differences in the community compositions among two groups. As shown in Fig. 3, a cladogram representation of significantly different taxa among groups was performed by LEfSe. The microbial composition was significantly different at the genus level, with 16 significantly different genera among the two groups. Acinetobacter, Ralstonia, Leptothrix, Sphingomonas, Ochrobactrum, Afipia, Leptotrichia, Oribacterium, and Moryella exhibited a relatively higher abundance in infected subjects, and can be considered H. pylori-enriched genera. Alloprevotella, Aggregatibacter, Klebsiella, Leptotrichlaceae__G_1_, Fusobacterium, Parvimonas, and Peptococcus were relatively more abundant in uninfected subjects, which could be considered to be decreased in the infected group. These higher or lower expressed genera in infected subjects can be considered as H. pylori-associated genera (LAD > 2, p < 0.05).

Fig. 3
figure 3

Comparison of microbial variations at the genus level, using the LEfSe online tool. a. Histogram of the LDA scores for differentially abundant features among groups. The threshold on the logarithmic LDA score for discriminative features was set to 2.0. N = uninfected group, P = infected group. b. Cladogram for taxonomic representation of significantly differences among groups. Differences are represented in the color of the most abundant taxa (red indicating P group, green indicating N group, and white indicating non-significant). N = uninfected group, P = infected group

Full size image

Eradication therapy for H. pylori partially changed salivary microbiota

To determine the effects of H. pylori eradication therapy on salivary microbial composition, saliva samples from a subgroup of H. pylori infected subjects (n = 22), were collected before (pre-eradicated group) and 2 months after treatment; saliva collected after successful eradication were classified into the eradicated group. The within-individual diversity in the samples from eradicated group was lower than pre-eradicated group (Additional file 7: Figure S3B). The Shannon diversity index, ACE richness index, and Chao 1 richness estimator were higher in pre-eradicated subjects than in eradicated subjects, with a significant difference between groups by t-test (Shannon p = 0.015, ACE p = 0.003, Chao 1 p = 0.002), indicating significant alteration in the within-individual diversity in samples from eradicated subjects compared to their baseline samples (Table 2).

Table 2 Alpha diversity indices for saliva bacteria in each group at 97% identity
Full size table

The beta diversity using unweighted UniFrac showed significant differences in the overall microbial composition between pre-eradicated and eradicated groups (PERMANOVAR, pseudo-F: 3.34, p = 0.001) (Fig. 2b). The NMDS also exhibited clear segregations in community structures among groups (Additional file 9:Figure S5B).

The relative difference of H. pylori-associated taxa was compared before and after eradication by LEFse analysis (Additional file 10:Figure S6). Among the H. pylori-enriched genera, Ralstonia, Leptotrichia, Sphingomonas, Leptothrix, Oribacterium, and Acinetobacter increased after the eradication, while Ochrobactrum decreased after the successful eradication (p < 0.05, paired Wilcoxon rank-sum test). Of the genera that decreased in infected subjects, Alloprevotella, Aggregatibacter, Leptotrichlaceae__G_1_, Parvimonas, and Fusobacterium decreased after the eradication (p < 0.05, paired Wilcoxon rank-sum test). Besides, we found that at phyla level, Fusobacteria increased after H. pylori eradication.

H. pylori in the oral cavity

H. pylori was detected in 38 out of 80 saliva samples, regardless of the H. pylori status in stomach, occupying 0.0139% of all the total sequences. (Fig. 4). 12 of the 34 subjects in infected subjects (35.3%), 11 of 24 subjects in the uninfected group (45.8%), and 15 of 22 subjects in the eradicated group (68.2%) were found to possess H. pylori in the oral cavity, respectively, (p = 0.054). The H. pylori signature was present in the saliva of subjects with negative 13C-UBT, RUT, and in H. pylori infected individuals after successful H. pylori eradication. We’ve also compared the prevalence of H. pylori both before and after H. pylori eradication. Pre-eradiation 7 subjects has positive saliva and 15 were negative. After successful eradication, H. pylori was still detected from saliva in 6 of the 7 positive subjects. Interestingly, 9 in 15 subjects who were H. pylori negative before eradication had it detected post H. pylori eradication whereas 6 remained free of H. pylori after eradication.

Fig. 4
figure 4

H. pylori in oral cavity of three groups. Red represent P group, yellow represent N group, and blue represent E group. N = uninfected group, P = infected group, E = eradicated group

Full size image

Discussion

In this study, using the technology of high-throughput sequencing, we found H. pylori exit in the oral cavity of infected (12/34), uninfected (11/24) and successful eradicated (15/22) subjects, composing 0.0139% of the total sequences.

A few studies have shown that H. pylori can be spread by oral-oral (or fecal-oral) way [2]. Feeding, kissing, and tableware shared play the important role in the transmission. Previous studies have detected the presence of H. pylori in saliva [3]. Bacteria in plaque adhere to the gums, which are relatively fixed. It is more reasonable to study the saliva instead of dental plaque in this situation. A comprehensive and thorough investigation of the bacterial diversity of salivary microbiota is essential for understanding the how or whether H. pylori infection alters the salivary microbiota. The technology of high-throughput sequencing has provided new cognizance of the structures and compositions of microbiota communities.

By comparing the alpha diversity indexes we found that the bacterial diversity in saliva was similar among the H. pylori uninfected and H. pylori infected people. Our study is consistent with the notion that H. pylori in the stomach has little or no effect on the bacterial diversity of the oral cavity [20]. The Shannon diversity index, ACE richness index, and Chao 1 richness estimator all declined after eradication of H. pylori compared to the baseline samples (p < 0.05), which was consistent with the prior studies that use of PPI and antibiotics may affect the oral microbiome [21, 22].

According to the beta diversity analysis based on the unweighted UniFrac distances, the community structures of saliva microbiota were different in H. pylori uninfected and infected individuals, which was contrary to the results of Schulz’s study [20]. Samples from the H. pylori infected subjects tended to cluster together, while the microbiota in the uninfected subjects appeared to be more variable suggesting that gastric H. pylori infection may affect oral bacterial components. Clear segregations by the PCoA and NMDS analysis among individuals before and after H. pylori eradication therapy demonstrated that successful eradication or eradication therapy changed the oral bacterial components to some extent.

In addition to the presence of different bacterial members, the abundance of some bacteria also differed significantly among groups. We clearly observed that some bacteria in the saliva of H. pylori infected individuals showed a significantly reduced abundance, among which Aggregatibacter, Klebsiella, Fusobacterium, and Parvimonas species were pathogenic bacteria, causing infective endocarditis, liver abscess, pneumonia, meningitis, systemic infections, et al. [23,24,25]. However, the abundance of other bacteria significantly increased in saliva of H. pylori infected individuals, most of which were oral microbiota composition, including Sphingomonas, Ochrobactrum, Afipia, Leptotrichia, Oribacterium, and Moryell, except Acinetobacter species causing infectious diseases like pneumonia and urinary tract infections [26], and Leptotrichia species, a potential cariogenic bacteria [27]. While in Schulz’s study, no significant difference in oral communities between H. pylori infected and uninfected individuals was detected at genus level [21], this may be due to the different target sequencing region of 16 s rDNA, sample size, or geographic location. Interestingly, most H. pylori-enriched genera increased after the eradication, including Ralstonia, Leptotrichia, Sphingomonas, Leptothrix, Oribacterium, and Acinetobacter. The exception was Ochrobactrum. However, genera low expressed in H. pylori infected saliva experienced a further decline after H. pylori eradication therapy, including Alloprevotella, Aggregatibacter, Leptotrichlaceae__G_1_, Parvimonas, and Fusobacterium. Our study suggests that H. pylori infection may change the salivary microbiota, eradication therapy would further change the bacteria composition of saliva microbiota. Although the clinical significance of these alterations was not known, H. pylori unexpectedly and clearly altered the oral microbiota composition.

In our study, we found that H. pylori infection didn’t change the abundance and diversity, but changed the composition structure of salivary microbiome. H. pylori is colonized in the stomach, and there is a certain distance from stomach to the oral cavity. Even though H. pylori can cause a large change in the abundance of the microbiota in the stomach, it is difficult to affect the abundance of the salivary microbial flora. Previous studies have reported acid inhibition in upper gastric tract may have an effect on the oral microbiome leading to alterations in the microbiota [28]. In addition, changes in gastric pH could also lead to an alteration in the pH of oral cavity [29]. We proposed that H. pylori likely changed the community and structure of oral microbiota through changes in the acidic environment in stomach by generating large amount of urease, an enzyme which decomposes urea into ammonia and carbon dioxide and transiently reduce the acidic environment in the stomach [30], leading to the further changes in pH of oral cavity. The use of PPIs during the eradication therapy would further inhibit the pH in stomach, leading to further alteration in saliva microbiota, which can partially explain why genera enriched in H. pylori infected individuals would further increase and genera low expressed in H. pylori infected individuals would decline after successful eradication. Although the precise mechanism has yet to be clarified, to our knowledge this is the first study to clearly show oral microbiota alterations as a result of H. pylori infection in a cohort of subjects. Additional studies like metagenomics sequencing or metabolomics to investigate these possible causal relationships would likely provide interesting findings.

Using amplicon sequencing of 16S rDNA V3-V4 hypervariable regions we detected H. pylori in the oral cavity of almost half of the subjects regardless of whether they had gastric infection with H. pylori. Subjects who did not have H. pylori in the oral cavity before eradication surprisingly had H. pylori detected in saliva samples after H. pylori eradication therapy. Clearly, using these techniques the prevalence of H. pylori in oral cavity is not clearly associated with colonization status in the stomach which is not consistent with the notion that the oral cavity represents a secondary site for H. pylori colonization [31]. The gastric and oral mucosa differ markedly. For example, of the two only the gastric mucosa expresses Lewisb antigen, an ABO blood group antigen, which enables adherence of H. pylori to the epithelial surfaces. It has been proposed that H. pylori is a passerby in oral cavity, rather than a colonizer and it may be also be included in the material in gastroesophageal reflux. The natural history of H. pylori infection has been that after H. pylori eradication from the stomach, gastric reinfection is rare and when it occurs early it can often be shown to be recrudescence (the same genotype) whereas later reinfections are most often reinfection with a different genotype [32]. The hypothesis that H. pylori was a common passerby rather than a colonizer would partly explain why recurrences are most common in areas with poor sanitation and a high prevalence of H. pylori and rare in developed countries whose frequency of H. pylori infection had become much lower than that of poor regions.

Strengths and limitations

Our study showed that H. pylori infection and the eradication treatment resulted in alterations of oral microbiota. However, there were limitations to our study. The technique of high-throughput sequencing we used in our study could not detect H. pylori-specific virulence factors such as VacA, CagA, OipA, etc. or full sequence [33]. Although the samples of the first set of cross-sectional analysis were collected at different times in August 2018, the collection interval between the first and the last collection were equal. All the samples underwent DNA extraction and sequencing analysis at the same time to make our result were comparable. One issue with the interpretation is that there was no control sample of H. pylori uninfected individuals receiving the same antimicrobial therapy which precluded determination about whether the presence of H. pylori, the antimicrobial therapy, or both were dominant factors in changing the within-individual diversity of the oral cavity.

Conclusions

Our study using high-throughput sequencing showed that H. pylori is present commonly in the oral cavity with no clear relation to H. pylori infection of the stomach. Both H. pylori infection and eradication therapy caused alterations in community and structure of the oral microbiota.